Using BM25 weighting and Cluster Shrinkage for Detecting Duplicate Bug Reports

Abstract

In software maintenance, bug reports play an important role for the correctness of software packages. Unfortunately, a duplicate bug report problem arises because there are significant many duplicate bug reports in various software projects. Processing duplicate bug reports is thus time-consuming and has high cost of software maintenance. In this research, we propose a detection scheme based on the BM25 weighting and cluster shrinkage (BM25-CS) to enhance the detection performance. The effectiveness of this method is verified in an empirical study with three opensource projects, SVN, Argo UML, and Apache. The experimental results show that our method outperforms other detection schemes about 6-10% in all cases.